Europe's first five-car autonomous race ends in collision as Kinetiz wins at Imola

What happened at Imola

On September 5, 2026, five fully autonomous race cars took the grid at Italy's Imola Circuit during the ACI Racing Weekend for a 12-lap contest. According to a report on dev.to, it was Europe's first five-car, wheel-to-wheel race between driverless vehicles, organised by the Abu Dhabi Autonomous Racing League, better known as A2RL. Kinetiz, a team from the UAE, took the win, with Constructor Racing of Germany second and Italy's PoliMOVE third. Unimore, also from Italy, finished fourth and Germany's TUM completed the field in fifth.

The running did not go cleanly. Dev.to's account describes Unimore recording the fastest lap of the event, a 1:40, before a technical problem stopped its car. PoliMOVE then collided with the stopped Unimore machine, and the race was ended at that point with Kinetiz still leading. The incident-filled finish is part of the story rather than a footnote: it shows exactly where these systems currently struggle.

Identical cars, different code

Every team competed with the same EAV-25 hardware, which shifts attention away from vehicle engineering and toward software. With a shared platform underneath, the result functions something like a controlled benchmark: each car faced the same physical problem, and the finishing order largely reflects how well each team's perception, planning and control stack handled it. The dev.to piece reads the event this way too, arguing that identical hardware made the teams' autonomous driving software the central point of comparison.

Why a race is harder than a lone run

A single driverless car circulating a track mainly has to follow the racing line and respect track limits. Adding four competitors introduces moving obstacles, overtaking decisions, close-quarters spacing and the possibility that another car will behave unpredictably, all resolved in fractions of a second. Imola produced both failure modes on the same day: a technical fault that parked a front-running car, and contact between two vehicles. Citing A2RL's official race report, dev.to notes that the event was the league's first held outside Abu Dhabi, and that racing offers a controlled but high-pressure setting for perception, planning and real-time decision-making.

Why it matters

The result is a genuine milestone for real-time AI: five independent software stacks were pushed into direct competition, under time pressure, in a shared physical space, and produced a meaningful ranking. It also illustrates a distinction the autonomous driving field keeps relearning, namely that peak performance and consistent performance are separate problems. Unimore was the quickest car over a single lap and finished fourth; Kinetiz was still running at the front when the race was called.

The dev.to report is careful about what the event does not show. A circuit is a closed, defined environment, and nothing from Imola speaks to public-road safety, deployment costs or integration with commercial fleets. The transferable takeaway is the engineering pattern itself: turning live sensor input into safe action quickly enough for the task, and handling exceptions when something breaks. That pattern reaches well beyond motorsport, into robotics, inspection systems and industrial automation, where the same questions about response time, input interpretation and human oversight apply.

What to watch next

The next indicators for A2RL will be whether future events continue to test multi-car competition rather than solo runs, how reliably teams manage incidents like the Imola collision, and whether the league publishes further evidence from its testing and simulation work. A first international race outside Abu Dhabi gives the programme a reference point; repeating the exercise without contact would mark the real progress.